Diversity of supernovae Ia determined using equivalent widths of Si II 4000

0.88 MB·English

by  V. Arsenijevic

Astronomy & Astrophysics manuscript no. diversitySnIa.ars.08.ii (cid:13)c ESO 2009
January 5, 2009
Diversity of supernovae Ia determined using equivalent widths of
Si II 4000
V. Arsenijevic, S. Fabbro, A. M. Mour˜ao, A. J. Rica da Silva
CENTRA - Centro Multidisciplinar deAstrof´ısica, IST, AvenidaRovisco Pais, 1049 Lisbon, Portugal
e-mail: [email protected]
9
0 Received 25 July 2008/ Accepted 17 September2008
0
ABSTRACT
2

n Aims. Spectroscopic and photometricproperties of low and high-zsupernovaeIa(SNeIa) havebeen analyzed in order
a to achievea better understandingof their diversity and toidentify possible SN Iasub-types.
J Methods. We use wavelet transformed spectra in which one can easily measure spectral features. We investigate the

5 Si II 4000 equivalent width (EWw{Si II}). The ability and, especially, the ease in extending the method to SNe at
high-z is demonstrated.

] Results. We applied the method to 110 SNe Ia and found correlations between EWw{Si II} and parameters related
h to the light-curve shape for 88 supernovae with available photometry. No evidence for evolution of EWw{Si II} with
p
redshiftisseen.Threesub-classesofSNeIawereconfirmedusinganindependentclusteranalysiswithonlylight-curve
o- shape, colour, and EWw{Si II}.
r Conclusions. SNe from high-z samples seem to follow a similar grouping to nearby objects. The EWw{Si II} value
t measured on a single spectrum may point towards SN Ia sub-classification, avoiding the need for expansion velocity
s
gradient calculations.
a
[
Key words.supernovae:general – methods: data analysis

2
v
3 1. Introduction the major concerns regarding the use of SNe Ia in cosmol-
3 ogy is a possible systematic difference between the low-z
1 The peak luminosities of type Ia supernovae (SNe Ia) and high-z samples (Blondin et al. 2006; Garavini et al.
3 are one of the best distance indicators at high redshifts. 2007a; Bronder et al. 2008). Reasons for such evolution
9. Consequently, due to the relatively small dispersion in are generally thought to be related to changes in metal-
0 their light-curves, they are used for cosmological param- licity or composition of the progenitor or circumstellar
8 eter estimation (e.g., Riess et al. 1998; Permutter et al. medium, progenitor mass or delay times. It is already
0 1999). However, some SNe seem to depart from standard known that progenitor age is a relevant issue for variabil-
v: behaviour and deserve further attention. The most repre- ity in SN peak luminosity (Prieto, Stanek & Beacom2008;
i sentative examples are the sub-luminous 1991bg-like ob- Gallagher et al. 2005) and that metallicity might evolve
X jects, the unusual SNe 2002ic and 2005gj1, and the even with redshift (see recent work of Ellis et al. 2008 and ref-
r moreoutstandingSN2003fg(Howell et al. 2006).Theex- erence therein). We shall demonstrate that the equivalent
a
istence of a single family of SNe Ia is thus still debatable width ofthe Si II 4000feature, EW {Si II} defined below,
w
(seeFilippenko 1997forareviewonSN(in)homogeneity). coupledwithlight-curveparameters,isanindicatorofpos-
Hence separatingsub-types of SNe Ia or attempting a con- siblesub-classeswithinSNe Ia,consistentforbothlowand
tinuous parametrization of spectra and/or light-curves in- high-z SNe.
cludingthefulldiversityofSNeIaobserveduptonowcould
reducethescatterintheHubblediagramandimprovetheir
2. The Si II 4000 spectral feature
use in cosmology.
In order to quantify the spectral differences between Theideabehindtheuseofspectralfeaturestostudysuper-
SNe Ia, Nugent et al. (1995) proposed a ratio between novae relies on the possibility of defining normalized spec-
the depths of absorption features of Si II at 5972˚A and tralratiosofthese featuresthatarecommonforallSNe.It
6355˚A. This ratio was also found to correlate with the ab- hasbeendemonstrated(Nugent et al. 1995;Folatelli2004;
solute magnitude of SNe Ia and the light-curve shape pa- Benetti et al. 2005;Bronder et al. 2008)thattheratiosof
rameter. We further investigate this idea exploring a con- the line depths or strengths are correlated with the light-
sistent method of measuring SN features on transformed curve shape. Here we analyze the possibility of extracting
spectra using wavelets as will be explained below. One of from spectral features additional information on intrinsic
SN properties and their potential to distinguish SN sub-
1 SomeauthorsarguethatSN2002ic maybeatypeIcsuper- classes at both low and high-z.
novainstead ofIa(e.g.Benetti et al. 2006;Wang et al. 2004), For instance, the Si II λ6355 absorption feature near
and the SN 2005gj classification as type Ia/IIn is still doubtful 6150˚Ais widely used,being the mostcharacteristicfeature
(Prieto et al. 2007; Trundleet al. 2008). of a SN Ia. Available high-z spectra however rarely extend
2 Arsenijevic et al. 2008: Diversity of SNeIadetermined using equivalent widthsof SiII 4000
to a rest-frame wavelength of 6150˚A, thus we are forced original spectrum scale 15
tporeffoercaubslyoutrheatfeteantutiroens Cona tIhIeHb&lKuerorpaSritIIof40t0h0e isnpeocrtdrear, 000001......024680 00000.....0011189012
to compare the nearby with the high-z sample (Folatelli 4000 4500 5000 5500 6000 6500 4000 4500 5000 5500 6000 6500
2004). l
The rest-framepeak flux atthe time ofmaximum lightoc- scale 14 scale 13
curs at about 4000˚A. At earlier epochs it is shifted slightly 000...234 000...011505
0.1 0.00
to lower wavelengths, while at later epochs the spectrum −00..01 −−00..1005
−0.2 −0.15
peaksatlongerwavelengths.Weareparticularlyinterested 4000 4500 5000 5500 6000 6500 4000 4500 5000 5500 6000 6500
in this spectral range since almost all high-z spectra ex-
hibit the Si II λ4130 feature blueshifted to 4000˚A, which scale 12 scale 11
mseiegmhsttsoerbveeawgeolloidncdhiositcinegfourischoimngpaSrNiseonIawfirtohmloSwN-zeSIbNea.nIdt −−−0000000.......00111100505505 −−−00000000........01234321
4000 4500 5000 5500 6000 6500 4000 4500 5000 5500 6000 6500
Ic when the usual Si II 6150 feature is not available, as it
is often the case with high-z SNe. The Si II 4000feature is scale 10 scale 9
mReodstdleynlioncgateeffdecbtestwsheoeunld39a0l0so˚Abaenddi4m20in0i˚Ashiendtshiencreestth-firsafmeae-. −−00000.....0011005005 −−−00000000........0011211005050505
ture is narrow.Consideringallthese interestingproperties, 4000 4500 5000 5500 6000 6500 4000 4500 5000 5500 6000 6500
the feature is a good candidate to be used for cosmology
(Bronder et al. 2008). scale 8 scale 7
0.2 0.10
The intensity of the Si II 4000 feature is however often 0.1 0.05
0.0 0.00
low and contamination can be important. To study this −0.1 −0.05
−0.2 −0.10
feature we apply a simple wavelet denoising procedure as 4000 4500 5000 5500 6000 6500 4000 4500 5000 5500 6000 6500
described below.
scale 6 scale 5
0.02 0.02
0.01 0.01
0.00 0.00
2.1. Discrete wavelet transform −0.01 −0.01
−0.02 −0.02
4000 4500 5000 5500 6000 6500 4000 4500 5000 5500 6000 6500
Wavelets are considered as a suitable tool for studying
local properties of a signal due to their specific structure. scale 4 scale 3
Itnodeperdo,cetshseitrhelodcaaltafraetqudeniffceyrenretprreesseonlutattioionns aolrlowscsaleuss. −−−0000000.......000000000111100505505 −−00000.....000000011005005
Therefore any local defect of a signal does not affect 4000 4500 5000 5500 6000 6500 4000 4500 5000 5500 6000 6500
the whole wavelet decomposition. A local inhomogeneity
in the SN spectrum, such as remaining residuals after scale 2 scale 1
sautmbtorsapchtieornicoafbsthorepthioonst tghaaltaxiys nspoetctcroummp,lestkeylyermemissoivoend, −−−0000000.......000000000000000246642 −−−0000000.......000000000000000123321
4000 4500 5000 5500 6000 6500 4000 4500 5000 5500 6000 6500
in the reduction process, or any calibration or instrumen-
tal undesirable effect, affects just the coefficients in the
waveletdecompositioninthesmallregionwhereitappears. Fig.1. Inversewavelet reconstruction of separate scales of the
wavelet decomposition of a spectrum of SN 1994D. The sum of
To provideahomogeneoussampleofdata,weuse spec- contributions from scales 7-14 is used to construct the wavelet
tra in the interval 3400˚A−7000˚A. Other choices of inter- transformed spectra, Fw(λ), used in thiswork.
vals do not make much difference due to the local charac-
ter of the wavelet transform. All spectra were deredshifted
first.Mallat’spyramidalgorithm(Mallat1989)isthenper-
formed to obtain the Discrete Time Wavelet Transform trum becomes much smoother,and one caneasily measure
(DTWT) using Daubechies’ extremal phase wavelets with the features.
4 vanishing moments (D8) as a basis. The analysis of the wavelet power spectrum2 revealed
The number of decomposition levels refers to the num- that scale 15 shows much greater discrepancy among SNe,
beroflevelsofsmootheddataandwaveletcoefficients.The even for spectroscopically similar events. For this reason,
coefficientsatthelowerlevelscarry“high-frequency”infor- scale15willnotbeincludedinthewaveletreconstructionof
mation. If a spectrum consists of N = 2m discrete points thespectra.Wecansupportthisstepbythefollowingtech-
(which can be assumed without loss of generality), there nical explanation:the contributionfrom scale 15,shownin
will be m wavelet coefficient bands and a scaling value, Fig. 1, is obtained by the inverse wavelet transform of a
whose inverses are given in the decomposition shown in scaling constant, equal to the sample mean multiplied by
Fig. 1 (see Nason & Silverman 1994). the square root of the number of original data points, and
Itisoftenreasonabletoassumethatonlyafewlargeco- the featureless inverse of the wavelet coefficient from the
efficientscontainrelevantinformationabouttheunderlying largest scale. The subtraction of the largest scale, as also
signal, while small wavelet coefficients, especially from low mentioned in Starck, Siebenmorgen & Gredel (1997), does
scales, can be attributed to noise or any other undesirable not significantly deform the features of the spectrum.
highfrequencyfeature.Thusthispartofspectra,thescales
1-6 (see Fig. 1), will be excluded from our considerations. 2 Obtainedbysummingthesquaresofthewaveletcoefficients,
This step assures that after taking the inverse DTWT of divided by the number of corresponding coefficients for each
such a spectrum, the resulting wavelet transformed spec- scale.
Arsenijevic et al. 2008: Diversity of SNeIadetermined using equivalent widthsof SiII 4000 3
After removal of scales 1-6 and scale 15 in the wavelet
space,weapplytheinverseDTWT.WetreatallSNeIathe
same way to obtain a homogeneous sample of normalized
wavelet transformed spectra. One could also apply a stan- lmin lmax
0.8
dard(soft)thresholding,butanoptimalchoiceforathresh-
oldvaluethatcutsoffallsmallerwaveletcoefficientsandis
consistentforwholevarietyofSNspectraisacomplexissue 0.6
andwehavenotattemptedits implementation.Mostspec-
tralfeaturesanalysesincludeaboxcarorGaussiansmooth-
Fw
ing. Residual sky lines or galactic emission lines are thus 0.4
alsosmoothedleadingtopossiblemisestimationsoffeature
boundsandextrema.Ourapproachgreatlyreducesthisef-
fect. 0.2
2.2. Determination of EWw{Si II} 0.0
In order to quantify SN Ia diversity with spectral features, 3400 3600 3800 4000 4200 4400
rather than looking at deviation from the average SN Ia l
spectrum (James et al. 2006), some authors use a pseudo-
equivalent-width (EW) (Hachinger, Mazzali & Benetti Fig.2. Illustration of the EWw{Si II} definition on a wavelet
2006; Garavini et al. 2007a; Bronder et al. 2008). Here, transformed spectrum,Fw(λ), of SN 1994D.
we apply the same definition but on the wavelet trans-
formed spectra:
Standarderrorpropagation,assuming smallenougher-
EW = N Fwc(λi)−Fw(λi)∆λ, (1) rors, applied on Eqn. 1 leads to:
w Fc(λ )
wprhoecreeduFrwe(λt)haistthhaestrXiba=en1esnforpmerefdworSmNiedfluoxnafStNer tshpeecwtraav,elaest σstat ="i=N1(cid:18)FσF2wcw2((λλii)) + FFwcw24((λλii))σc2(λi)(cid:19)(∆λ)2#1/2,
described in 2.1 (therefore the subscript w); Fc(λ) stands X
w
for the pseudo-continuum of the transformed flux, N is where σFw consists of the inverse DTWT of the low scale
the number of data points; the bounds of the feature are coefficients fromthe waveletdecompositionthathavebeen
λmin=λ1 and λmax=λN. subtracted; σc(λ) is obtained from the error of a straight
line fit that defines the pseudo-continuum.
EW {SiII}isdefinedastheEW ofSiIIλ4130feature
w
on transformed spectra, When we have multi-epoch spectra, EWw{Si II} is
estimated for each of the epochs, and the error-weighted
EW {Si II}≡EW {Si IIλ4130}.
w w polynomial fit returns the propagated σstat. In the more
To determine the pseudo-continuum bounds, we use a difficult but common cases of single epoch spectra, we
multipleGaussianpeak-fittingsemi-automatedroutine.All add to σstat an error floor computed as the maximum
the pseudo-EW definitions used in the literature, includ- dispersion observed on the epoch of the given spectrum.
ing ours, depend on the accuracy of the pseudo-continuum Typically we add σfloor =1.7.
bounds, and therefore on the spectrum signal-to-noise ra-
tio (S/N). The traditional definition of EW does allevi- The dispersion seen in wavelengths of the feature
ate the dependency on S/N, but requires the full spectrum boundscanbepartlyexplainedbyredshiftuncertainty,typ-
continuum to be well defined, which is more difficult for ically of 3-30 ˚A (respectively for ∆z of 0.001 and 0.01) at
the SNe Ia.We arecurrentlyinvestigatingthissubject.We a redshift of 0.5. However, no offset in λmin or λmax was
elaborateontheS/NdependencyinSect.2.3.Wealsonote foundbetweengoodqualityspectraandreducedS/Ndata.
that our wavelet-based method for the removal of high- The major reason for this is the use of the specific wavelet
frequencyfeaturesaddsextrarobustnesstothepseudo-EW transformwe performonSN spectra to obtainthe spectral
estimation. function on which EWws are measured. The main conse-
The EW {Si II} evolves smoothly with the SN epoch, quence is that statisticaluncertainties inthe high-z setare
w
as we verified on a spectral series template given by the generally higher due to lower S/N.
SALT2 model (Guy et al. 2007). We use the same pre-
scription as Bongard et al. (2006) to obtain an estimate
2.3. Systematic errors
ofEW {SiII}att=0atmaximuminB:forasingleepoch
w
spectrum, we use the actual value (but increase the error), To test the robustness of the EW {Si II} measurements
w
fortwoepochspectraastraightlinefit,andformulti-epoch and quantify various systematic errors, we ran a set of
spectra a quadratic polynomial fit. simulations taking very high S/N low-z spectra on top
We also estimate the statistical error made by the de- of which we added extra contributions (for instance sky
termination of extrema of the specific features considered, and/or galaxy) and Gaussian noise with a variable stan-
as both the wavelet transformed flux and the pseudo- dard deviation.
continuum come with uncertainties that affect the calcu- One source of systematic errors is the host galaxy con-
lated value of EW . tamination.Veryoften,thelightfromthehostgalaxyisnot
w
4 Arsenijevic et al. 2008: Diversity of SNeIadetermined using equivalent widthsof SiII 4000
perfectly removed and the remaining light affects the esti- Source Size
mation of equivalent widths. We tested different amounts statistical error 15−27% (low-z vshigh-z)
of galaxy contamination,up to 80% of the total integrated galaxy contamination <20−40% (Evs S type)
SN flux from 3400-7000˚A. The simulations using template low S/N <5% for S/N ≥5
galaxy spectra of Hubble types E and Sa indicate that reddening <0.5%
increasing galaxy contamination implies lower values of high frequency feature residuals <0.5%
the measured EW {Si II}, as found in Garavini et al.
w
(2007a); Bronder et al. (2008). In Fig. 3 we illustrate the Table 1.SummaryoftheerrorbudgetfortheEW {SiII}
w
correlation between increasing galaxy contamination level measurements.
and the relative decrease in EW {Si II}.
w
correctionneverexceeds0.5%,whileforEW ofCaIIH&K
w
reaches the order of 1%. This fact supports the choice of
0.6
E galaxy theSiII4000featureinsteadofbroaderones,suchasCaII
Sa galaxy
H&K or Mg II 4300. Other extinction laws based on stud-
ies of dust properties in the MagellanicClouds givesimilar
0.4 results for the wavelength range under consideration (see,
e.g. Pei 1992; Weingartner & Draine 2001) and thus were
Si II} not further investigated at this stage.
EW{w The contamination from residual sky line subtraction
Si II} / 0.2 wasalsostudied.Toachievethis,weaddedtooursetofsev-
W{w erallow-zSNe,whichweredshiftedatz =0.5,afiducialsky
E
D spectrumandGaussiannoise.WeverifiedthatourDTWT
0.0 procedureproperlyremovestheselines,leavingEWw{SiII}
unmodified.
In Table 1 we summarize all the measurement errors
mentioned above.
−0.2
0.0 0.2 0.4 0.6 0.8
q
3. Data sets
Fig.3. Relative error of the measured EW {Si II} as a
w
function of q, the amount ofhost contaminationfor ellipti- We then applied our measurement procedure to a set of
cal (open symbols) and spiral (filled symbols) galaxies for published SNIa data.The mainselectioncriterionwasthe
3 low-z SNe, 1981B, 1994D and 2002bo (given as triangle, detection of the Si II spectral feature we are interested in.
diamond and square symbols respectively). It thus depends on wavelength and epoch coverage. Tests
performedonSNspectraltemplates indicatethatitshould
be present within an interval of [−15,15] days relative to
B-maximum. All selected SN spectra have epochs within
OurresultswithafewSNeshowascatterthatisgener- an interval of ±9 days around the maximum in B in the
ally of less than 20% for SNe in elliptical galaxies,even for rest-frame.
contaminations of 50-80%,although it becomes greater for
Furthermore, if the presence of strong host galaxy con-
objects hosted in spiral galaxies,but never exceeds 40% of
taminationwasnoticedresultinginseveremisestimationof
thetruevalue.Itmayhappen,however,thatextremelyhigh
Si II 4000 feature, that spectrum was rejected. We mainly
host-galaxycontaminationdistortssignificantlythefeature
selected SNe with published photometry in order to check
wemeasure,asinthecaseoftheSN2002bospectrumwith
for correlations between EW {Si II} and light-curve pa-
80%ellipticalgalaxycontamination(seeFig.3).Recallthat w
rameters.
Bronder et al. (2008) make a selection cut when galaxy
contamination is found to be more than 65% since it can
produce a change of up to 100% in the EW value. 3.1. Low-z sample
We expect that at low signal-to-noiseratios,spectra do
notexhibitsoclearlytheSiII4000feature,makingitdiffi- The analysis is applied to a sample of 35 local SNe taken
culttomeasureaccuratelyandpossiblyintroducingaslight from the literature and presented in Table 2. The spec-
misestimation. Varying the noise in our simulations then troscopic data are taken mainly from public archives:
measuring EW {Si II} and corresponding error on each SUSPECT3, CfA Supernova Archive4 and SUSPEND5.
w
spectrum suggests that no significant bias is present in the Spectra of 1991T and 1991bg-like SNe, also of other pe-
EWw{Si II} values. culiar SN events were also considered. EWw{Si II} is mea-
sured on 124 spectra of low-z SNe.
We also checked our hypothesis of the small effect of
reddeningontheSiII4000feature.Furthersimulationsus-
ingtheextinctionlawofCardelli, Clayton & Mathis(1989) 3 http://bruford.nhn.ou.edu/∼suspect/index1.html
with R = 3.1 have shown that the relative error of 4 http://cfa-www.harvard.edu/oir/Research/supernova/SNarchive.html
V
EW {Si II} measured on spectra with/without reddening 5 http://www.nhn.ou.edu/∼jeffery/astro/sne/spectra/spectra.html
w
Arsenijevic et al. 2008: Diversity of SNeIadetermined using equivalent widthsof SiII 4000 5
3.2. High-z sample Benetti et al. (2005) using velocity gradients inferred
from the measured blueshift of the Si II absorption fea-
The principal difficulty when studying high-z SNe, besides ture at 6355˚A. The authors identified sub-classes of SNe
much noisier spectra compared to the low-z events, is the
in the low redshift sample termed: high velocity gra-
wavelength coverage spanned by spectral observations re-
dient (HVG), low velocity gradient (LVG) and FAINT
sultinginthelackofmajorspectralfeatures,likeSiII6150.
SNe. The latter group includes those SNe that are
However this is not the case with the Si II 4000 feature.
found to have similar brightness as that of SN 1991bg.
From the available set of high-z SuperNova Legacy
Other authors confirmed the same sub-classifications
Survey (SNLS) data, we used 26 SNe with spectral epochs
(see for instance Hachinger, Mazzali & Benetti 2006;
within ±9 days relative to maximum in B that have been
Pastorello et al. 2007b).Theseauthorsalsousetheexpan-
published in Howell et al. (2005). This SN sample is dis-
sion velocity gradient of the Si II λ6355 feature to classify
tributed over0.337<z <1.01 (see Table 4). Usually there
SNe, therefore multi-epoch spectra are required. A similar
is only one spectrum for each high-z supernova. We note
grouping was found by Branch et al. (2006) studying the
that the host galaxy spectra were not subtracted.
equivalent widths of Si II λ6355 and λ5972features. These
High-zspectrafromthefirsttwoyearsoftheESSENCE
approaches however are hardly applicable to high-z SNe
project (see, e.g., Matheson et al. 2005; Miknaitis et al.
due to the wavelength coverage.
2007)werechecked;amongthese,26spectrathatshowthe
Wedistinguishedexactlythesamethreesub-classesap-
presence of the Si II 4000 feature were used. No attempts
plying a hierarchicalcluster analysis in a much smaller pa-
to subtract the host galaxy from the spectra were made,
rameterspace,usingonlyEW {SiII}withaSALT2light-
but Matheson et al. (2005)didnotreportanystrongcon- w
curve shape, x1, and colour parameter.
tamination within this selected sample.
Weemphasizeasanadvantageofourmethodtheability
The Supernova Cosmology Project (SCP) high-z spec-
to identify the 3 sub-classes with only one spectrum rela-
tra used in this work were provided by Hook et al.
tively close to B-maximum. If we remove the EW {Si II}
(2005) and Lidman et al. (2005), as seen from Table 4. w
parameterfrom the cluster analysiswe are not able to find
Measurements were performed on 6 and 10 SN spectra re-
the three clusters of SNe. We also ran the cluster analy-
spectively.
sis on low-z SNe using other estimates of light-curvewidth
We use also 7 spectra from High-z Supernova Search
such as ∆m15, stretch, ∆(MLCS) and identified the same
Team(HzSST)thatwerepublishedin Tonry et al. (2003).
three sub-classes.
Spectrafrom Hook et al. (2005)and Tonry et al. (2003)
It is known that for HVG6 SNe the Si II λ6355 line
were corrected for host galaxy light.
evolves rapidly (Benetti et al. 2005). These SNe generally
have larger photospheric velocities than SNe with a slower
Light-curve data come from Astier et al. (2006);
evolution of the Si II λ6355 feature, such as the LVG SNe.
Miknaitis et al. (2007); Permutter et al. (1999);
This latter group includes both normal SNe Ia and the
Tonry et al. (2003); Kowalskiet al. (2008).
brightestones.Again,intheclassificationof Branch et al.
(2006), the LVG group corresponds to core-normals and
4. EWw{Si II} properties shallows. However, SNe from both HVG and LVG groups
have similar maximum luminosities. The difference in pho-
Armed with EWw{Si II} values for 32 low-z and 75 high-z tospheric velocity comes as a consequence of the difference
SNe Ia, we searched for correlations with light-curve pa- in photospheric temperature; the HVG have a lower tem-
rameters. In order to apply a consistent procedure for all perature compared to LVG SNe.
SNe, we fitted the available light-curves with the SALT2
TherearealsoafewSNe,namely1983G,1984A,2002bf
model. We ended up with 30 low-z and 58 high-z SNe, for
and 2004dt, that show similar behaviour to that of SN
whichEW {SiII}andSALT2parameterswerecalculated,
w 2002bo,awellstudiedHVGsupernova.Thesimilaritiesbe-
as can be seen in Tables 3 and 4. We also tried our anal-
tween these SNe were also pointed out by Altavilla et al.
ysis with published parameter values from the MLCS2k2
(2007) and Leonard et al. (2005). All of them belong
(Jha, Riess & Kirshner2007) andSALT (Guy et al. 2005;
to the group of SNe with unusually high photospheric
Kowalskiet al. 2008) fitters.
velocities. In addition, 2004dt and 2002bf are both highly
polarized. Spectropolarimetry can provide, in general, a
4.1. Relation with the light-curve parameters probe of supernova geometry; greater divergence from
spherical symmetry normally causes a higher polarization,
We first checked the correlation between EWw{Si II} and but the latter may also be caused by clumping (see
x1 (SALT2 width parameter), illustrated in the top panel Wang & Wheeler 2008for the latestreview).Furthermore,
of Fig. 4. The x1 parameter measures the departure of the Wang, Baade & Patat (2007) found significant peculiarity
stretchofa SN fromthe averagevalue ofthe trainingsam- of SN 2004dt comparing the degree of polarization across
ple(instandarddeviations);itsaveragevalueisadoptedto the Si II λ6355 line and light-curve decline parameter
satisfy < x1 >= 0 and < x21 >= 1 (Guy et al. 2007). ∆m15.
Nugent et al. (1995) and Hachinger, Mazzali & Benetti
(2006) found a strong correlation between the ratio of the The difference between HVG and LVG SNe can also
depthoftheabsorptionfeaturesofSiIIat5972and6355˚A, be studied in objects that have similar decline rates, as
R(Si II), and EW{Si IIλ5972} and ∆m15. We confirm a shown in Tanaka et al. 2008 for SNe 2002bo and 2001el.
similar correlation for low and high-z SNe using the Si II TheyconcludethatburninginLVGislesspowerfulthanin
4000 feature and a different light-curve model (SALT2).
We noticed however several SNe departing from the 6 This group corresponds to the broad-line group in the clas-
main trend. A similar behaviour was highlighted by sification of Branch et al. (2006).
6 Arsenijevic et al. 2008: Diversity of SNeIadetermined using equivalent widthsof SiII 4000
(2006); Hachinger, Mazzali & Benetti (2006), namely SNe
1989B,1991M,1992Aand2004eo.Amongthem,SN1991M
is labeled as HVG SN, while the others are shown as LVG
HVG SNe objects. We may add to this list an HVG SN 2002er that
40 LFVAGIN TS NSeNe has been found to have properties common to both HVG
SNLS SNe
ESSENCE SNe and LVG SNe, like a lower expansion velocity but temper-
SCP SNe
HzSST SNe atureshigherthanofotherHVGSNe (Benetti et al. 2005;
30 92A 02bf Tanaka et al. 2008).Thesefeweventscanbeconsideredas
91M 81B 02bo transitional objects, linking all three sub-classes and pro-
Si II} 94D 02er viding continuitybetweenthe groups.We mayalsoinclude
EW{w 20 04eo89B 03cg99ac dSiNca1te9s9;4iDn ianddthiteiosen,oSbNjec2ts00a5shkthseeetmopsptoaneesltainbliFsihg.a4liinnk-
96X
05cf
05am 90N between the LVG and FAINT SNe. Similar plots like in
10 99by 00cx98aq98bu040S3du99ee Fstirgo.n4g aoffif nEiWtywo{fStihiIsI}suvpeerrsnuosva∆fmor15thoerFsAtrIeNtcThgcroonufipr.m a
00E 06gz
97cn AnotherobjectofinterestisthepeculiarSN2006gz(anSN
86G 05hk 99aa
with the slowest fading light-curves ever seen in a SN Ia),
91T
0 99aw whosepropertiesdeviate fromthe LVGSNe.Its early-time
Si II velocity is low, like for LVG SNe, which is attributed
−4 −2 0 2 4
to an envelope of unburned carbon that slows expansion
x1 (SALT2) (see Hicken et al. 2007).
A linear fit in upper plot in Fig. 4, with only LVG SNe
at low-z included, gives:
HVG SNe
LVG SNe
40 FSANILNST SSNNee EWw{Si II}=(12.21±0.32)−(5.88±0.30)x1.
ESSENCE SNe
SCP SNe
HzSST SNe
92A
30 02bf
91M
81B 02bo
EW{Si II}w 20 94D 04e0o2er 89B 03cg
96X E/S0 galaxies
99ac Spiral galaxies
05cf 90N 98bu Peculiar galaxies
10 980a3qdu 00cx04S05am99ee 99by Pec 86G 03du
06gz 00E
0 −0.5 99aa909a.0w97cn91cTo0l5ohrk (SA0L.T52) 816.G0 1.5 Morphological type SabSbcScd 91T 99aa05hk 04S000E6g9z8aq9999beye 98bu 90N99ac 04e8o9B 02er 8012Bbo910M024bdft
S0 00cx 05am 94D 92A
Fig.4. EWw{Si II} versus x1 (top) and colour (bottom) esti- E2
mated using SALT2 including high-z SNe. The colours chosen
97cn
fordifferentsub-classesarethesameasin Benetti et al. (2005); E0 96X LHVVGG SSNNee
Pastorello et al. (2007b):bluefilledsquaresforHVGSNe,light FAINT SNe
greenfilledcirclesforLVGSNeandredsymbolsforFAINTSNe.
0 5 10 15 20 25 30
EWw{Si II}
Fig.5. Morphological type of the SN host galaxy versus mea-
sured EWw{Si II} for nearby SNe.
HVGSNe,thusthereisadifferenceinkineticenergyofthe
ejecta(ofthe orderof∼2%)7.Thisdifference inkinetic en-
ergy implies that HVG SNe light-curves are narrowerthan
LVG SNe ones. Even if we assume that it is the only dif- We looked for the SN host galaxy type to ex-
ference between these two groups of SNe, this might cause plore its effect on the SN sub-class. It was empha-
an intrinsic dispersion in the luminosity/light-curve shape sized by Hamuy et al. (1996); Altavilla et al. (2004);
relation. Gallagher et al. (2005)thatintrinsicallyfainteventsrather
As shown in Fig. 4, we also find 4 events considered as occurinE/S0galaxies.ThediagraminFig.5pointstothe
non-standard in Pastorello et al. (2007b); Branch et al. conclusion that early-type (E/S0) galaxies lack SNe with
smaller EW {Si II} values, say ≤ 12. Among our sample,
w
7 Although this differenceis too small to explain thespectral noHVGSNhostedintheellipticalgalaxyisfound,though
diversity between theHVGand LVG SNe. therearetwoambiguousSNebetweenLVGandHVGgroup
Arsenijevic et al. 2008: Diversity of SNeIadetermined using equivalent widthsof SiII 4000 7
in S0 galaxies. These two are identified as SNe 1992A and results confirm the possible contamination of the high-z
1994D. sample with HVG SNe. This fact deserves further atten-
Regarding the issue of whether or not the distance es- tion as a potential source of systematic error, as was also
timator varies acrossdifferent environments,Conley et al. suggested in Tanaka et al. 2008.
(2006) suggest that stretch correction works well no mat- Previouscomparisonsbetweenlowandhigh-zSNespec-
ter the SN environment. On the other hand, it is not the tral features have shown the homogeneity within the two
case with colour, since this correction accounts for dust or samples.Wehavetakenastepfurtherbyintroducinganew
extinction effects and also for the intrinsic SN relationship approachthatallowsusto identify SNIasub-classes,inde-
(see, e.g. Sullivan et al. 2006; Conley et al. 2007). pendent of the SN redshift, using only one spectrum with
Si II 4000 together with the light-curve width and colour
parameters.
4.2. High-z SNe
Themethodpresentedinthis workallowsustoexaminein
5. Conclusions
aconsistentwaylowandhigh-zSNeforwhichwecalculated
EWw{Si II} and SALT2 light-curve parameters. The use of SNe Ia for cosmology relies on empirical cali-
Adding high-z SNe to the cluster analysis one obtains bration techniques on the light-curves and K-correction.It
the following SNe that are classified as HVG SNe: 03D4fd, wouldbereassuringtohaveausableindicatormeasureddi-
04D3dd, 04D3gx, b013 and d058. rectly fromSN spectrathatcanbe usedasanindependent
Two outliers with the largest x1 values can also be no- calibrator of the luminosity of an SN Ia event. Applying
ticed in upper panel of Fig. 4, namely SCP SN 1997S and a straightforward transformation using wavelets, we were
ESSENCE d033. In the opposite region of lower x1 values, able to estimate, in a consistent manner on a fair num-
besides SN e138, there are a few candidates, SCP 1997ai ber of SNe Ia, the equivalent width of the Si II 4000 fea-
SNLS-03D4cz,03D4cnandESSENCEf216,e029thatseem ture which previously showed potential use for cosmology.
to have affinities towards the group of FAINT SNe. At low redshift, we were able to automatically distinguish
ESSENCE SN d117 is identified in the ambiguous zone three classes of SNe Ia previously found by other authors
betweenFAINTandLVGSNe.Further,itishardtoclassify usingmuchmoreinformation,suchasthe expansionveloc-
SNe 03D1fq, 03D4fd, e132 and f011. ity gradient.
We do not exclude the possibility that dispersion seen The same wavelet-based approach was applied to high
inupperplotinFig.4isduetohostgalaxycontamination, redshiftdata,revealingananalogousclusteringtendencyto
especially for the SNLS and ESSENCE SNe. As it is men- that found for low redshift supernovae. Yet it is not clear
tioned in2.3 the presence ofgalaxylightin the SN spectra whether the three sub-classes, found in the low-z and ver-
might lead to underestimated values of EWw{Si II}. ified in the high-z sample, are completely distinct or come
The bottom panel in Fig. 4 indicates that the cor- fromonecontinuousfamily.Ifso,thiscouldpossiblybethe
relation between EWw{Si II} and c is quite weak, but remnant of an extra parameter or a more complex mod-
there are outliers.It canjustify aneventualcolourcut, say elling of the supernova data. The implementation of our
−0.2 < c < 0.2, that would ensure a more homogeneous method using recently available much larger spectral sam-
SN data set for cosmological use. Indeed, SNe with c out ples andthe use of different SN Ia sub-classesin cosmolog-
of the mentioned interval tend to have more dispersion in ical analysis is now in preparation.
theHubblediagram(see Arsenijevic2008;Kowalskiet al. Although the method presented here is promising and
2008). we are planning to apply it to other spectral features,
In addition, we find no significant correlation between we remark that line ratios, pseudo-equivalent widths and
EWw{Si II} and the absolute magnitude MB corrected theircurrentderivativesarenotoptimaltoextractinforma-
for x1 and c, using updated values for α and β from tiononlowersignal-to-noisespectra.Parameterscalculated
Guy et al. 2007,althoughthere aremanyoutlierspresent. from spectral features, such as equivalent widths, still con-
The linear factor in the relation MBcorr ∼ EWw{Si II} sideronlylocalinformationandeachfeatureindependently.
for the low-z SNe is of the order of 0.01. We used New indicatorsbasedoncombininglocalinformation,such
values of the distance modulus µ for low-z SNe from as wavelet coefficients from the whole spectra, can provide
Tammann, Sandage & Reindl (2008). These questions are a better characterisationof the supernovae.
further discussed in Arsenijevic (2008).
A source of systematic errors in the Hubble diagram Acknowledgements. This work was supported by Fundac¸a˜o para
comes from uncertainties in distance or host galaxy a Ciˆencia e Tecnologia (FCT), Portugal under POCTI/CTE-
AST/57664/2004. V. Arsenijevic acknowledges support from FCT
reddeningorfromintrinsicpropertiesoftheSNe.Withthe
under grant no. SFRH/BD/11119/2002 and S. Fabbro grant no.
available data we are not yet able to distinguish between SFRH/BPD/31817/2006. Most of the code was written in R8. We
potential systematic effects that come from magnitude wouldliketothankG.Nasonforusefulcomments ontheimplemen-
dependence on intrinsic colour and the colour contribution tationofwavethreshpackage.
due to dust effects in the host galaxy of the SNe. Note
that in order to examine the dust properties of distant
References
galaxies using SNe Ia, one needs to first remove the
intrinsic dependence of SNe luminosity on the light-curve Altavilla,G.etal.2007, A&A,475,585A
parameters. Altavilla,G.etal.2004, MNRAS,349,1344
8 R Development Core Team (2007). R: A language and
An important issue in cosmologywith SNe Ia is how to environment for statistical computing. R Foundation for
clean the SN sample of any objects that might not obey a Statistical Computing, Vienna, Austria. ISBN 3-900051-07-0,
standard peak luminosity/light-curve shape relation. Our URL http://www.r-project.org
8 Arsenijevic et al. 2008: Diversity of SNeIadetermined using equivalent widthsof SiII 4000
Anupama,G.C.,Sahu,D.K.&Jose,J.2005,A&A,429,667 Starck,J.-L.,Siebenmorgen, R.&Gredel,R.1997,ApJ,482,1011
Arsenijevic, V. 2008, PhD Thesis (in preparation), Technical Strolger,L.G.2002,AJ,124,2905
UniversityofLisbon Sullivan,M.etal.2006,ApJ,648,868
Astier,P.etal.2006, A&A,447,31 Tammann, G., A., Sandage, A. & Reindl, B. 2008, arXiv:astro-
Barbon,R.etal.1990,A&A,237,79 ph/0806.3018
Benetti,S.etal.2006,ApJ,653,L129 Tanaka, M.etal.2008,ApJ,677,448
Benetti,S.etal.2005,ApJ,623,1010 Tonry,J.L.etal.2003, ApJ,594,1
Benetti,S.etal.2004,MNRAS,348,261 Trundle,C.etal.2008,A&A,483,47
Blondin,S.etal.2006,AJ,131,1648 Turatto,M.etal.1998, AJ,116,2431
Bongard,S.etal.2006, ApJ,647,513B Turatto,M.etal.1996, MNRAS,283,1
Branch,D.etal.2006,PASP,118,560 Valentini,G.etal.2003, AJ,595,779
Branch,D.etal.2004,PASP,116,903 Wang, L. & Wheeler, C. J. 2008, Annual Review of Astronomy and
Branch,D.etal.2003,AJ,126,1489 Astrophysics,46,433
Branch,D.etal.1983,ApJ,270,123 Wang,L.,Baade,D.&Patat,F.2007,Science, 315,212
Bronder,J.etal.2008,A&A,477,717 Wang,L.etal.2004,ApJ,604,L53
Brown,P.J.etal.2005,ApJ,635,1192 Weingartner,J.C.&Draine,B.T.2001, ApJ,548,296
Cardelli,J.A.,Clayton,G.C.,&Mathis,J.S.1989,ApJ,345,245 Wells,L.A.etal.1994,AJ,108,2233
Conley,A.etal.2007,ApJ,664,L13
Conley,A.etal.2006,ApJ,132,1707
Elias-Rosa,N.etal.2006,MNRAS,369,1880
Ellis,R.S.etal.2008, ApJ,674,51
Filippenko, A. V. 1997, Annual Review of Astronomy and
Astrophysics35,309
Folatelli,G.2004, NewAstronomyReviews48,623
Gallagher,J.S.etal.2005,ApJ,634,210
Garavini,G.etal.2007a, A&A,470,411
Garavini,G.etal.2007b,A&A,471,527
Garavini,G.etal.2005,AJ,130,2278
Garavini,G.etal.2004,AJ,128,387
Garnavich,P.M.etal.2004,ApJ,613,1120
Gomez,G.&Lopez, R.1998,AJ,115,1096
Guy,J.etal.2007,A&A,466,11
Guy,J.etal.2005,A&A,443,781
Hachinger,S.,Mazzali,P.A.&Benetti,S.2006, MNRAS,370,299
Hamuy,M.etal.2002,AJ,124,417
Hamuy,M.etal.1996,AJ,112,2391
Hicken,M.etal.2007,ApJ,669,L17
Hook,I.M.etal.2005, AJ,130,2788
Howell,D.A.,Sullivan,M.,Nugent,P.etal.2006,Nature,443,309
Howell,D.A.etal.2005,ApJ,634,1190
James,J.Berianetal.2006,MNRAS,370,933
Jeffery,D.J.etal.1992,ApJ,397,304
Jha,S.,Riess,A.G.,Kirshner,R.P.2007,ApJ,659,122
Jha,S.etal.1999, ApJS,125,73
Kirshner,R.P.etal.1993,ApJ,415,589
Kotak,R.etal.2005,A&A,436,1021
Kowalski,M.etal.2008,accepted inApJ,arXiv:astro-ph/0804.4142
Krisciunas,K.etal.2007,AJ,133,58
Leonard,D.C.etal.2005, ApJ,632,450
Li,W.etal.2003,PASP,115,453
Li,W.etal.2001,PASP,113,1178
Li,W.etal.1999,AJ,117,2709
Lidman,C.etal.2005,A&A,430,843
Mallat,S.G.1989,IEEETrans.PAMI,11,674
Matheson,T.etal.2005,AJ,129,2352
Mazzali,P.A.,Danziger,I.J.&Turatto,M.1995,A&A,297,509
Mazzali,P.A.etal.1993,A&A,269,423
Miknaitis,G.etal.2007, ApJ,666,674
Nason, G. P. & Silverman, B. W. 1994, The Discrete Wavelet
Transform inS,J.Comp.Graph.Statist.3,163
Nugent,P.etal.1995,ApJ,455,L147
Pastorello,A.etal.2007a,MNRAS,376,1301
Pastorello,A.etal.2007b,MNRAS,377,1531
Patat, F.etal.1996,MNRAS,278,111
Pei,Y.C.1992,ApJ,395,130
Perlmutter,S.etal.1999,ApJ,517,565
Phillips,M.M.etal.2007,PASP,119,360
Phillips,M.M.etal.2006,AJ,131,2615
Phillips,M.M.etal.1992,AJ,103,1632
Phillips,M.M.etal.1987,PASP,99,592
Pignata,G.etal.2004,MNRAS,355,178
Prieto,J.L.,Stanek, K.Z.,&Beacom, J.F.2008,ApJ,673,999
Prieto,J.L.etal.2007,arXiv:astro-ph/0706.4088
Quimby,R.,Ho¨flich,P.&Wheeler,C.J.2007,ApJ,666,1083
Riess,A.G.etal.1998, AJ,116,1009
Sahu,S.etal.2008,ApJ,680,580
Salvo,M.E.etal.2001,MNRAS,321,254
Stanishev,V.etal.2007,A&A,469,645
Arsenijevic et al. 2008: Diversity of SNeIadetermined using equivalent widthsof SiII 4000 9
SN name Epoch Bands References
(with respect to B maximum)
1981B 0 UBV 1
1986G -3, -1, 0, 1, 2 BV 2
1989B -5, -1, 0, 3, 5 UBVRI 3, 4
1990N 2, 7 UBVRI 5
1991M 3 VRI 6
1992A -5, -1, 3, 5 UBVRI 7
1994D -5, -4, -2, 2, 4 UBVRI 8
1996X -2, 0, 1 BVRI 9
1998aq 0, 1, 2, 3, 4, 5 UBVRI 10
1998bu -3, -2, -1 UBVRI 11
1999ee -4, -2, 0, 5 UBVRI 12
2000E -3, -2 UBVRI 13
2002bf 3 BVRI 14
2002bo -5, -4, -3, -2, -1, 4 UBVRI 15
2002er -5, -4, -3, -2, -1, 0, 2, 4, 5 UBVRI 16, 17
2003cg -5, -1, 1, 4 UBVRI 18
2003du -4, -2, -1, 0, 1, 3, 4 UBVR 19, 20
2004dt -5, -4, -3, -2, -1, 0, 1, 2, 3, 4 ··· 14, 21
2004eo -3, 2 UBVRI 22
2004S 1 UBVRI 23
2005am -1, 2 UBV 24
2005cf -4, -3, -1, 0, 4, 5 UBVRI 25, 26
2005hj -5, 0, 2, 3, 5 ··· 27
2006gz -4, -1 UBVri 28
SNe 91bg-like
1991bg 1, 2 BVRI 29
1999by -5, -4, -3, 3, 4, 5 UBVRI 30
SNe 91T-like
1991T -3, 0 UBVRI 31, 32, 33
2000cx -3, -2, 0, 1, 5 UBVRI 34
SNe Ia peculiar
1997br -4 UBVRI 35
1997cn 3 UBVRI 36
1999aa -3,-1 UBVRI 37
1999ac 0, 2 UBVRI 38, 39
1999aw 3, 5 BVRI 40
2002cx -4, -1 BVRI 41, 42
2005hk -5, -4, 4 UBVRI 43, 44
Table 2. List of nearby SNeused in this work.
References: (1) Branch et al. 1983; (2) Phillips et al. 1987; (3)
Barbon et al. 1990; (4) Wells et al. 1994; (5) Mazzali et al. 1993; (6)
Gomez & Lopez 1998; (7) Kirshner et al. 1993; (8) Patat et al. 1996;
(9)Salvo et al. 2001;(10)Branch et al. 2003;(11)Jha et al. 1999;(12)
Hamuy et al. 2002;(13)Valentini et al. 2003;(14)Leonard et al. 2005;
(15) Benetti et al. 2004; (16) Pignata et al. 2004; (17) Kotak et al.
2005; (18) Elias-Rosa et al. 2006; (19) Anupama,Sahu & Jose
2005; (20) Stanishev et al. 2007; (21) Altavilla et al. 2007; (22)
Pastorello et al. 2007b; (23) Krisciunas et al. 2007; (24) Brown et al.
2005; (25) Garavini et al. 2007b; (26) Pastorello et al. 2007a; (27)
Quimby,H¨oflich & Wheeler 2007; (28) Hicken et al. 2007; (29)
Turatto et al. 1996; (30) Garnavich et al. 2004; (31) Jeffery et al.
1992; (32) Phillips et al. 1992; (33) Mazzali, Danziger & Turatto 1995;
(34) Li et al. 2001; (35) Li et al. 1999; (36) Turatto et al. 1998; (37)
Garavini et al. 2004; (38) Garavini et al. 2005; (39) Phillips et al.
2006; (40) Strolger et al. 2002; (41) Li et al. 2003; (42) Branch et al.
2004; (43) Phillips et al. 2007; (44) Sahu et al. 2008
10 Arsenijevic et al. 2008: Diversity of SNeIadetermined using equivalent widthsof SiII 4000
SN name zCMB mB x1 c EWw{Si II}
1981B 0.0072 12.0074±0.0217 −0.6660±0.1673 0.1530±0.0196 27.47±2.33
1986G 0.0028 12.0184±0.0371 −3.0859±0.3261 0.9851±0.0409 3.56±2.67
1989B 0.0036 12.2424±0.0177 −1.0569±0.1418 0.4736±0.0141 20.67±1.48
1990N 0.0045 12.6599±0.0138 0.7208±0.1168 0.0773±0.0121 15.68±2.45
1991M 0.0076 14.3553±0.1376 −1.5904±0.3325 0.0036±0.1140 28.53±2.31
1992A 0.0059 12.5356±0.0100 −2.0636±0.0995 0.0810±0.0103 30.37±2.42
1994D 0.0027 11.7314±0.0095 −2.1674±0.0842 −0.0643±0.0100 23.62±2.47
1996X 0.0078 12.9832±0.0151 −1.2519±0.1539 0.0408±0.0153 17.86±2.61
1998aq 0.0042 12.2893±0.0070 −0.4507±0.0615 −0.0920±0.0076 9.69±2.52
1998bu 0.0042 12.0777±0.0062 −0.3570±0.0556 0.3320±0.0068 12.95±1.37
1999ee 0.0105 14.8268±0.0062 0.4171±0.0526 0.3052±0.0067 10.91±2.09
2000E 0.0048 12.8034±0.0099 0.4851±0.1117 0.2058±0.0101 8.13±2.47
2002bf 0.0247 16.2978±0.0445 −0.4205±0.2040 0.2293±0.0275 29.29±2.32
2002bo 0.0053 13.9533±0.0086 −0.6595±0.0668 0.4711±0.0090 26.85±2.35
2002er 0.0086 14.2263±0.0100 −1.1062±0.0825 0.1918±0.0102 23.18±2.32
2003cg 0.0041 15.7912±0.0089 −0.4629±0.0655 1.3005±0.0091 20.69±2.40
2003du 0.0064 13.4923±0.0093 0.1035±0.0697 −0.0472±0.0087 11.33±0.94
2004dt 0.0188 ··· ··· ··· 29.48±2.35
2004eo 0.0147 15.0554±0.0066 −1.4892±0.0601 0.1191±0.0075 20.03±4.19
2004S 0.0091 14.1485±0.0293 −0.3168±0.1622 0.1317±0.0166 8.30±2.58
2005am 0.0090 13.7303±0.0225 −2.6235±0.1490 0.2102±0.0167 13.40±3.09
2005cf 0.0070 13.0841±0.0054 −0.3598±0.0471 −0.0387±0.0060 14.70±1.53
2005hj 0.0574 ··· ··· ··· 8.36±2.76
2006gz 0.0277 15.7957±0.0176 2.3344±0.1741 0.0490±0.0148 7.60±4.17
SNe 91bg-like
1991bg 0.0042 14.6012±0.0381 −2.7372±0.1940 0.7185±0.0309 ···
1999by 0.0029 13.6486±0.0143 −3.0586±0.2231 0.5483±0.0148 10.80±4.38
SNe 91T-like
1991T 0.0070 11.5450±0.0118 1.0398±0.1138 0.2008±0.0121 0.54±1.77
2000cx 0.0079 13.0444±0.0082 −1.3500±0.0790 0.0604±0.0090 13.12±2.92
SNe Ia peculiar
1997br 0.0080 13.4941±0.0160 0.0209±0.1319 0.2849±0.0128 ···
1997cn 0.0170 15.9641±0.1168 −2.6791±0.3456 0.1092±0.0753 4.57±3.24
1999aa 0.0144 14.7077±0.0131 1.1371±0.1313 −0.0242±0.0124 3.19±2.66
1999ac 0.0099 14.1078±0.0092 −0.1542±0.1064 0.1118±0.0102 17.33±2.35
1999aw 0.0393 16.7145±0.0163 2.0927±0.1799 −0.0069±0.0157 1.08±1.54
2002cx 0.0250 17.7281±0.0189 0.1253±0.1507 0.3253±0.0154 ···
2005hk 0.0118 15.9015±0.0112 −0.4887±0.1149 0.3112±0.0111 5.13±2.53
Table3.SALT2estimatesandEWw{SiII}resultsfornearbySNe.The
blank spaces “···” in EWw{Si II} column indicate insufficient presence
of SiII feature despite an adequate wavelength coverage.
SN name zCMB mB x1 c EWw{Si II} Reference
03D1ax 0.496 22.9691±0.0160 −1.0045±0.1715 −0.0304±0.0285 9.01±3.95 1
03D1bk 0.865 ··· ··· ··· 13.50±5.36 1
03D1co 0.68 24.1069±0.0481 0.9602±0.6335 −0.0229±0.0539 8.21±5.94 1
03D1ew 0.868 24.3489±0.0576 0.4624±0.4097 −0.0762±0.1830 9.94±8.66 1
03D1fq 0.80 24.5299±0.0343 −1.8585±0.5215 −0.2695±0.1905 31.30±4.38 1
Continued on next page...